In your noisy image/array you now have values which are outside of the original range [0 ; 1], which is incorrect and unrealistic as a pixel on your real-world 8 bit camera sensor cannot have negative values or values above 255. This means that you need to implement => all negative values in my noisy array are 0, all positive values in my noisy array that are higher than 1 are 1. However, this introduces the problem that you are effectively truncating your Gaussian distribution, meaning that what you added to the image as noise (Gaussian) is not what you get in the end (some strange malformed/truncated Gaussian).
Hello, thank you for pointing this out. I still need to understand what we should do to the pixel where the values are outside of the original pixel range since the textbook did not explain this problem.
Thank you bro!! Highly appreciated
You are welcome😄
In your noisy image/array you now have values which are outside of the original range [0 ; 1], which is incorrect and unrealistic as a pixel on your real-world 8 bit camera sensor cannot have negative values or values above 255. This means that you need to implement => all negative values in my noisy array are 0, all positive values in my noisy array that are higher than 1 are 1. However, this introduces the problem that you are effectively truncating your Gaussian distribution, meaning that what you added to the image as noise (Gaussian) is not what you get in the end (some strange malformed/truncated Gaussian).
Hello, thank you for pointing this out. I still need to understand what we should do to the pixel where the values are outside of the original pixel range since the textbook did not explain this problem.
1:00 I don't get what z stands for
z is a random number. It means that the Gaussian distribution takes any random number from negative infinity to positive infinity.
Thanks, that helped me to understand better :)
No worries :)
Thank you very much! It helped me a lot.
You are very welcome!
Very Helpful tutorial. Now the question is that how should i denoise the noised image???
Hi Asad, you can use the median filter to denoise the image. OpenCV has the median filter built-in function that you can use to denoise the image.